U.S. equities advanced, led by strong technology and semiconductor stocks, with the Philadelphia Semiconductor Index outperforming. While lower oil prices over the past week eased energy inflation concerns, the Fed raised rates by 25 bps as expected, potentially weighing on market momentum. The latest rate projections also indicate at least one more hike by year-end, as persistent inflation and resilient economic and labor-market conditions may warrant further tightening.
U.S. September flash manufacturing and services PMIs, released on September 23, both beat expectations and reached relatively high levels in recent years. New orders, employment and cost pressures also rose, prompting markets to price in further Fed tightening and higher-for-longer rates. Treasury yields initially eased alongside lower oil prices early in the week, but rebounded on continued hawkish Fed signals and strong PMI data. The U.S. 10Y Treasury yield broke above 5%, reaching 5.1% on the morning of September 24, weighing on bonds broadly.
As AI compute demand continues to surge, Google, Amazon, Meta, Microsoft and OpenAI are accelerating custom chip (ASIC) development to improve compute efficiency and lower costs. Citing industry data, we estimate ASIC shipments will surge 92% YoY to 14.46mn units in 2027, surpassing GPU shipments of 11.29mn units for the first time, signaling a shift toward a more diversified AI chip market.
Around 4.7mn industrial robots were operating in factories worldwide in 2024, triple the level a decade ago. Manufacturing powerhouses such as South Korea, China and Japan widely deploy robots across autos, electronics and metal processing, helping sustain cost and quality advantages and strengthening competitiveness. While industrial robot adoption in the U.S. still trails major Asian manufacturing economies, recent efforts to rebuild industrial capacity, reshore critical supply chains and expand semiconductor and pharmaceutical production could gradually reshape its industrial and manufacturing strategy.
Humanoid robots consist primarily of a “brain” combining semiconductors and AI models, and a body powered by sensors, actuators, wiring, connectors and lightweight batteries, enabling them to perceive their environment, reason and execute coordinated physical actions. Investment banks estimate actuators will account for over 50% of total humanoid robot costs by 2030, making them the highest-value core component. McKinsey even sees actuators and sensors as potential key supply-chain bottlenecks.
